Machine learning in medical emergencies: a systematic review and analysis

Abstract

Despite the increasing demand for artifcial intelligence research in medicine, the functionalities of his methods in health emergency remain unclear. Therefore, the authors have conducted this systematic review and a global overview study which aims to identify, analyse, and evaluate the research available on diferent platforms, and its implementations in healthcare emergencies. The methodology applied for the identifcation and selection of the scientifc studies and the diferent applications consist of two methods. On the one hand, the PRISMA methodology was carried out in Google Scholar, IEEE Xplore, PubMed ScienceDirect, and Scopus. On the other hand, a review of commercial applications found in the best-known commercial platforms (Android and iOS). A total of 20 studies were included in this review. Most of the included studies were of clinical decisions (n=4, 20%) or medical services or emergency services (n=4, 20%). Only 2 were focused on m-health (n=2, 10%). On the other hand, 12 apps were chosen for full testing on dif ferent devices. These apps dealt with pre-hospital medical care (n=3, 25%) or clinical decision support (n=3, 25%). In total, half of these apps are based on machine learning based on natural language processing. Machine learning is increasingly applicable to healthcare and ofers solutions to improve the efciency and quality of healthcare. With the emergence of mobile health devices and applications that can use data and assess a patient's real-time health, machine learning is a growing trend in the healthcare industry.